Research on Vehicle Model Risk Rating Based on GLM Model and K-Means Clustering Algorithm for Car Insurance Pricing Scenario

Under the background transformation in the automobile and insurance industries and diversified and personalized consumer demands, precise pricing of car insurance has become an important trend. In China, the vehicle model risk rating evaluation is still in the initial stage, there are still problems such as limited car-related factors, limited coverage of models, imperfect risk rating systems, etc. This paper studies the relevant theories of vehicle model risk rating at home and abroad, takes the claim data and vehicle configuration data as the basis, uses the GLM model based on gamma distribution to select the most influential factors, and carries out the risk evaluation for all fuel vehicle models in the market. K-means clustering algorithm is innovatively used to divide the vehicle risk into 30 levels, which provides an important reference for the construction of vehicle model risk rating system in China.